[address-policy-wg] 2006-05 New Policy Proposal (PI Assignment Size)

Michael,

On Sep 20, 2006, at 2:17 AM, Michael.Dillon at btradianz.com wrote:
>> Simply, PI to network topological leaves doesn't scale.
>
> On the contrary, PI *DOES* scale and it has scaled because
> there is no direct one-to-one connection between issuing a
> PI assignment and consuming a global routing table slot.

Yes, pedantically, if you don't insert a prefix into the routing  
system, it scales quite well.  I'm not sure this distinction has much  
value though.

> Those things have greater impact
> on global routing table size than the number of new PI
> blocks.

More specific announcement for TE may have a greater impact _today_  
but given the pattern of liberalization of PI policies in all the  
RIRs, it isn't clear to me this will be the case in the future.   
What's worse is that more specifics can be filtered while still  
potentially providing routability (albeit perhaps sub-optimally)  
through the supernet.  If you apply filters that affect PI prefixes,  
those prefixes are simply gone.

Rgds,
-drc